| 摘要 |
PROBLEM TO BE SOLVED: To recognize each document by respectively applying a document recognition method suitable to each document even if a recognition object is an object obtained by collecting a plurality of kinds of documents as one set.SOLUTION: Generation means 101 reads a document to generate image data representing the document. Recognition means 102 recognizes the kind of the document represented by the image when the generation means 101 generates image data. The recognition means 102 recognizes a kind corresponding to a method that succeeds in recognition of the kind of a first document from an image of the first document among a plurality of kinds of image recognition methods as the kind of the first document. The recognition means 102 uses the image recognition method corresponding to the recognized kind of the first document to recognize a kind of a second document from an image of the second document. Output means 103 outputs information representing an arrangement state of a document group necessary for an application on the basis of a result recognized by the recognition means 102.SELECTED DRAWING: Figure 2 |
